Investigate water in MD trajectories
Switch branches/tags
Nothing to show
Clone or download
Latest commit 87a4dea May 7, 2015
Failed to load latest commit information.
.gitignore Update gitignore Mar 11, 2015


Investigate water in MD trajectories.


Install using python install

This package was written and tested on Python 3.4. Python 2 is supported through the python-future package. Let me know if you run into any issues.

Using with MSMBuilder

Setup will install new subcommands to the msmb command.

  • SolventShellsFeaturizer: Featurize trajectories. Use it like any other MSMBuilder featurizer.

  • SolventShellsAssigner: Used for generating "assignments" of atoms to shells for visualization

  • SolventApplyComponents: We can leverage the linear coefficients of decomposition models like PCA or tICA to enhance visualization of "important" solvent atoms. This command takes as input the result of SolventShellsAssigner (which assigns solvent atoms to features) as well as a serialized decomposition model (PCA or tICA) and weights solvent atoms by the coefficients of the features to which they are assigned.

  • SolventWriteVMD: This will export data from the above command in a format readable by VMD. This command will also generate a script which loads the data into the user field in VMD. This allows, e.g., coloring and selection based on solvent tICA coefficients.

Method Overview

Shells schematic

"Solvent" atoms are binned into shells surrounding "solute" atoms. The count of atoms in each shells can be used as a feature vector for use with MSMBuilder. By assigning solvent atoms to specific features, by identifying important features we can actually identify individual, relevant solvent atoms. For example, in the figure above, the bright-red "solvent" atom is contained within two shells (features) identified as important (perhaps by tICA). Using the provided scripts, one can easily perform such a visualization in VMD.